1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-31 16:02:52 +01:00
Commit Graph

14 Commits

Author SHA1 Message Date
Evan Cheng
1ce02d180e Enable machine sinking critical edge splitting. e.g.
define double @foo(double %x, double %y, i1 %c) nounwind {
  %a = fdiv double %x, 3.2
  %z = select i1 %c, double %a, double %y
  ret double %z
}

Was:
_foo:
        divsd   LCPI0_0(%rip), %xmm0
        testb   $1, %dil
        jne     LBB0_2
        movaps  %xmm1, %xmm0
LBB0_2:
        ret

Now:
_foo:
        testb   $1, %dil
        je      LBB0_2
        divsd   LCPI0_0(%rip), %xmm0
        ret
LBB0_2:
        movaps  %xmm1, %xmm0
        ret

This avoids the divsd when early exit is taken.
rdar://8454886

llvm-svn: 114372
2010-09-20 22:52:00 +00:00
Bill Wendling
077afde4bf Machine sink could potentially sink instructions into a block where the physical
registers it defines then interfere with an existing preg live range.

For instance, if we had something like these machine instructions:

BB#0
  ... = imul ... EFLAGS<imp-def,dead>
  test ..., EFLAGS<imp-def>
  jcc BB#2 EFLAGS<imp-use>

BB#1
  ... ; fallthrough to BB#2

BB#2
  ... ; No code that defines EFLAGS
  jcc ... EFLAGS<imp-use>

Machine sink will come along, see that imul implicitly defines EFLAGS, but
because it's "dead", it assumes that it can move imul into BB#2. But when it
does, imul's "dead" imp-def of EFLAGS is raised from the dead (a zombie) and
messes up the condition code for the jump (and pretty much anything else which
relies upon it being correct).

The solution is to know which pregs are live going into a basic block. However,
that information isn't calculated at this point. Nor does the LiveVariables pass
take into account non-allocatable physical registers. In lieu of this, we do a
*very* conservative pass through the basic block to determine if a preg is live
coming out of it.

llvm-svn: 105387
2010-06-03 07:54:20 +00:00
Dan Gohman
5736cd1e47 Start function numbering at 0.
llvm-svn: 101638
2010-04-17 16:29:15 +00:00
Evan Cheng
04b9deff58 Rever 96389 and 96990. They are causing some miscompilation that I do not fully understand.
llvm-svn: 97782
2010-03-05 03:08:23 +00:00
Evan Cheng
ee44d6a752 Look for SSE and instructions of this form: (and x, (build_vector c1,c2,c3,c4)).
If there exists a use of a build_vector that's the bitwise complement of the mask,
then transform the node to
(and (xor x, (build_vector -1,-1,-1,-1)), (build_vector ~c1,~c2,~c3,~c4)).

Since this transformation is only useful when 1) the given build_vector will
become a load from constpool, and 2) (and (xor x -1), y) matches to a single
instruction, I decided this is appropriate as a x86 specific transformation.
rdar://7323335

llvm-svn: 96389
2010-02-16 21:09:44 +00:00
Dan Gohman
44e25ed254 Don't enable the post-RA scheduler on x86 except at -O3. In its
current form, it is too expensive in compile time.

llvm-svn: 90781
2009-12-07 19:04:31 +00:00
Dan Gohman
d5dbd3f588 Add a target triple so that this test behaves consistently across hosts.
llvm-svn: 85640
2009-10-31 00:15:28 +00:00
Dan Gohman
ad6c6a3d33 Fix MachineLICM to use the correct virtual register class when
unfolding loads for hoisting.  getOpcodeAfterMemoryUnfold returns the
opcode of the original operation without the load, not the load
itself, MachineLICM needs to know the operand index in order to get
the correct register class. Extend getOpcodeAfterMemoryUnfold to
return this information.

llvm-svn: 85622
2009-10-30 22:18:41 +00:00
Dan Gohman
076a3b5e25 Teach MachineLICM to unfold loads from constant memory from
otherwise unhoistable instructions in order to allow the loads
to be hoisted.

llvm-svn: 85364
2009-10-28 03:21:57 +00:00
Dan Gohman
00c9f3e905 Mark dead physregdefs dead immediately. This helps MachineSink and
MachineLICM and other things which run before LiveVariables is run.

llvm-svn: 85360
2009-10-28 01:13:53 +00:00
Evan Cheng
eb82c3743a Turn on post-alloc scheduling for x86.
llvm-svn: 84431
2009-10-18 19:57:27 +00:00
Dan Gohman
3c46142f12 Add a CHECK line to check the position of the second divsd.
llvm-svn: 83009
2009-09-28 20:48:39 +00:00
Evan Cheng
ddc8678b00 Coalescer should not delete extract_subreg, insert_subreg, and subreg_to_reg of
physical registers. This is especially critical for the later two since they
start the live interval of a super-register. e.g.
%DO<def> = INSERT_SUBREG %D0<undef>, %S0<kill>, 1
If this instruction is eliminated, the register scavenger will not be happy as
D0 is not defined previously.
This fixes PR5055.

llvm-svn: 82968
2009-09-28 05:28:43 +00:00
Dan Gohman
487152872b Unbreak MachineLICM for instructions that reference RIP on x86-64 too.
llvm-svn: 82825
2009-09-25 23:58:45 +00:00